Browse Prior Art Database

Design for a distributed voice platform based on resource scheduling

IP.com Disclosure Number: IPCOM000005124D
Publication Date: 2001-Aug-15
Document File: 7 page(s) / 235K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a design for a distributed voice platform based on resource scheduling. Benefits include scalability, flexibility, and robust functionality in a distributed environment.

This text was extracted from a WORD97 document.
This is the abbreviated version, containing approximately 26% of the total text.

Design for a distributed voice platform based on resource scheduling

Disclosed is a design for a distributed voice platform based on resource scheduling. Benefits include scalability, flexibility, and robust functionality in a distributed environment.

General description

The disclosed design is for a distributed VoiceXML platform based on resource scheduling (see Figure 1). Service providers include:

Telephony Interface (TI)

Speech Recognizer (SR)

Text to Speech (TTS)

Dialogue Management (DM)

All service providers run on separate machines. Multiple backups may exist for each service provider, which is often necessary for a large-scale call center or voice portal that servers thousands of users (or more) at the same time.

The network layer provides the network connection for all the service providers through some kind of network protocol (such as TCP).

The resource scheduler is built on the network layer. When the resource scheduler receives requests from the VoiceXML interpreter and VoiceXML interpreter context, the scheduler allocates the services provided by the TI, SR, TTS, and DM components. The allocation is determined by using the resource scheduling table.

Figure 1 contains the block diagram of the distributed VoiceXML platform architecture.

Detailed Description

Resource scheduling table

The resource scheduling table records the current status of all the service providers in the distributed VoiceXML platform. At the request of the VoiceXML interpreter and the interpreter context, the resource scheduler allocates the related services based on the resource scheduling table.

An example of a resource scheduling table is contained in Figure 2. The svr_type column contains the service type of the server, which can be SR, TTS, DM, etc. The net_addr column contains the network address of the server. The port column contains the service port of the server. The max_svr_cap column contains the service capacity of the server, that is, the maximum capacity that the server can provide. The used_svr column contains the used number of the services in the server.

In the resource scheduling table, each row describes one server. These servers can provide different service types, and run on different machines or run on the same machine but monitor different ports. The servers may provide more than one service at the same time. An example of a resource scheduling table is contained in Figure 2. The second row in the table describes a TTS server, which can provide at most three TTS services at the same time, and two of them are in use.

Resource scheduler

The resource scheduler is the bridge between the distributed service providers and the VoiceXML interpreter and the VoiceXML interpreter context. It supports the dynamic configuration of the service providers (servers), the servers can be added into the distributed ...